Original Research ARTICLE
The diagnostic scope of sensor-based gait analysis in atypical Parkinsonism: further observations
- 1Department of Molecular Neurology, University Hospital Erlangen, Germany
- 2University Hospital Erlangen, Germany
- 3Medizinische Fakultät, Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany
- 4Department of Neurology, Medical University of Innsbruck, Austria
- 5Machine Learning and Data Analytics Lab, Germany
- 6Technische Fakultät, Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany
Background: Differentiating idiopathic Parkinson´s disease (IPD) from atypical parkinsonian disorders (APD) is challenging, especially in early disease stages. Postural instability and gait difficulty (PIGD) are core features of IPD and APD. Clinical evidence implies that patients with APD have larger PIGD impairment than IPD patients. Sensor-based gait analysis as instrumented bedside test revealed more gait deficits in APD compared to IPD. However, the diagnostic value of instrumented bedside tests compared to clinical assessments in differentiating APD from IPD patients have not been evaluated so far.
Objective: The objectives were a) to evaluate whether sensor-based gait parameters provide additional information to validated clinical scores in differentiating APD from matched IPD patients, and b) to investigate if objective, instrumented gait assessments have comparable discriminative power to clinical scores.
Methods: In a previous study we have recorded instrumented gait parameters in patients with APD (Multiple System Atrophy and Progressive Supranuclear Palsy). Here, we compared gait parameters to those of retrospectively pairwise disease duration-, age-, and gender-matched IPD patients in order to address this new research questions. To this aim, the PIGD score was calculated as sum of the MDS-UPDRS-3-items “gait”, “postural stability”, “arising from chair”, and “posture”. Gait characteristics were evaluated in standardized gait tests using an instrumented, sensor-based gait analysis system mounted on the lateral heel part of both shoes. Patients performed overground gait tests in self-selected walking speed. Machine learning algorithms were used to extract spatio-temporal gait parameters from raw sensor signals. Receiver Operating Characteristic analysis was performed in order to detect the discriminative power of the instrumented versus the clinical bedside tests in differentiating IPD from APD.
Results: A high Area Under the Curve (AUC) was shown by PIGD score (0.919), and UPDRS-3 (0.848). Particularly, the objective parameters stance time variability (0.841), swing time variability (0.834), stride time variability (0.821), and stride length variability (0.804) reached high AUC’s as well.
Conclusions: PIGD symptoms showed high discriminative power in differentiating IPD from APD supporting gait disorders as substantial diagnostic target. Sensor-based gait variability parameters provide metric, objective added value and serve as complementary outcomes supporting clinical diagnostics and long-term home-monitoring concepts.
Keywords: Neurologic gait disorders, Postural Balance, Biosensors, Parkinson Disease, Progressive supranuclear palsy (PSP), multisystem atrophy (MSA)
Received: 14 Nov 2018;
Accepted: 03 Jan 2019.
Edited by:Pedro J. Garcia-Ruiz, Hospital Universitario Fundación Jiménez Díaz, Spain
Reviewed by:Juan C. Martinez Castrillo, Hospital Universitario Ramón y Cajal, Spain
Raul Martinez Fernandez, Centro Integral en Neurociencias A.C. HM CINAC, Spain
Copyright: © 2019 Gassner, Raccagni, Eskofier, Klucken and Wenning. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Dr. Heiko Gassner, Department of Molecular Neurology, University Hospital Erlangen, Erlangen, 91054, Bavaria, Germany, firstname.lastname@example.org
Dr. Cecilia Raccagni, Department of Neurology, Medical University of Innsbruck, Innsbruck, Tyrol, Austria, Cecilia.email@example.com